normal Arg and Lys Exclusions for Nonbonded Interaction

  • mason
  • mason's Avatar Topic Author
  • Offline
  • Fresh Boarder
More
2 years 2 months ago #9296 by mason
Hello,

It looks to me like a protein topology generated by Martinize2 lacks nonbonded exclusions between the BB and SC2 atoms of Arg and Lys. Is this correct?

For some background:
I am using Martinize2 to generate a structure/topology for simulation in Gromacs 2021. I have this question after looking into exclusions as a means of validating my Martinize2 output. I am doing this so that if anyone ever asks me if I'm confident in my Martinize2 structure I can say "yes" and then tell them why.

Martinize2 makes a main topology with this directive:

[ moleculetype ]
molecule_0 1


That second digit is the value of "nrexcl" which Gromacs uses to exclude pairwise nonbonded interactions between neighbors. A value of 1 means only bonded neighbors will be excluded for nonbonded calculations.

A value of 1 is enough to exclude nonbonded interactions within most of the aminos in Martini3, since most residues only have 1 or 0 sidechain atoms. His, Phe, Trp, and Tyr all have explicit exclusions populated as the [ exclusions ] topology directive which fully disconnect atoms within those residues. I think this is thanks to the related directives in aminoacids.ff.

That leaves Arg and Lys, each of which has 3 atoms assembled as BB-SC1-SC2. The "nrexcl 1" referenced above will exclude the nonbonded interactions of BB with SC1, and SC1 with SC2, but not BB with SC2.

If I'm right, I'm wondering if this is an oversight or if it is intentional. If I'm wrong, I'll be very grateful to learn why. Thanks for reading!

Best,

Mason

Please Log in or Create an account to join the conversation.

  • mason
  • mason's Avatar Topic Author
  • Offline
  • Fresh Boarder
More
2 years 1 month ago #9306 by mason
If anyone thinks a different Topic is the better place for this I'll be glad to know.

Please Log in or Create an account to join the conversation.

More
2 years 1 month ago #9307 by riccardo
Hi,

I think this is a good place for this question. Sorry, but I missed it.

I quote some text from onlinelibrary.wiley.com/doi/pdf/10.1002/adts.202100391 (page 9):

"Finally, exclusions are applied between the two beads of the hinge that are not connected by any bond. It is worth noting that Martini does not have a defined exclusion rule in its parametrization protocol. In particular, although by default in Martini non-bonded exclusions are defined only between directly connected beads (i.e., only 1–2 interactions are excluded), for rigid fragments or molecules nonbonded exclusions are commonly applied between all the beads of the rigid fragment/molecule (see, e.g., the case of naphthalene just discussed). However, in general, nonbonded interactions are important for more flexible fragments or molecules, where keeping nonbonded interactions between beads separated by 2 (i.e., 1–3 interactions) or 3 (i.e., 1–4 interactions) bonds may be beneficial to capture nontrivial features of the interactions."

So, in general, Martini default is to use nrexcl=1 and there's no need to exclude nonbonded between BB and all the atom side chain beads. So the side chains that contain several beads *arranged in a linear fashion* like Arg and Lys will interact with BB beads. The fact that Trp and Tyr contain extra exclusions (under the Gromacs [ exclusions ] itp section) is because they are described by 4/5 beads all belonging to a planar structure and therefore it makes sense to exclude the interactions between them. (For the (so-called "hinge") construction of Trp, you can find some detailed info on Figure 3 (and associated text) of onlinelibrary.wiley.com/doi/pdf/10.1002/adts.202100391 . ) But I don't think interactions are excluded between the BB bead and the Trp beads not directly connected with that BB bead. I'm not sure Phe and His do actually have extra exclusions.

I hope this helps and feel free to follow up.

Please Log in or Create an account to join the conversation.

Time to create page: 0.092 seconds